Re: Help with 7 Circuit Project?

I guess that depends on the alarm dealer. Everyone I've ever discussed it with agrees that daytime break-ins are the most common. Most thieves want to get in and out with your stuff quickly and without confrontation. Most homes are unoccupied during the middle of the day on weekdays so that's when most burglaries occur.

Regardless what some imprisoned thief may have said, the fact is that most burglars avoid homes with alarms. There are more homes without alarms than with them and there's no reason to add to the risk by hitting a house with an alarm.

Dogs are indeed a good deterrent, probably even better than alarms. They make lots of noise and the thief doesn't know if the dog is going to bite. OTOH, alarms don't require walking, feeding, veteranarian visits, etc. :^)

Sounds like a description of me doing the samba -- something you really wouldn't want to watch. :^)

There are several things you can do to reduce the likelihood of false alarms. Selection, placement and programming of motion detectors should be at or close to the top of your list. If there will be free roaming pets in the home while the system is armed, the detectors need to be "pet resistant" type. These employ various schemes to differentiate between 2- and 4-legged targets.

The simplest "pet proof" detectors are aimed across the protected space at a level above the height of one's pets. This may work with some dogs but cats can jump from the floor to the top of a refrigerator. Another sensor type uses a split transducer to pick up IR (body heat) from an intruder. The split transducer receives IR from low and high level objects. A processor looks for motion in both upper and lower areas at the same time. If the motion is only along the floor or on top of the couch nothing happens. If there is simultaneous activity from both planes the detector throws a relay and the alarm system is triggered. There are other types of pet resistant technology but this is the most common method.

Dual technology detectors have microwave and passive infrared (PIR) sensors in the same housing. The alarm is not activated unless both sensors are triggered together. Because the detectors respond to different stimuli, there is lower risk of a false from a dual tech sensor than a single technology device.

As mentioned above, placement is an important aspect of false alarm prevention. During takeovers of competitors installations I've found countless instances where single tech PIR motion detectors were installed on inside walls facing directly at windows. This can be a major source of problems. PIRs will often trip if sunlight, car headlights or any other strong source of IR light is momentarily reflected through a window onto the surface of the detector. Many detectors' installation instructions clearly state that they should not face a window. Unfortunately, because interior walls are usually uninsulated, lazy (or poorly trained) techs like to place them there. A little extra effort is required to place the sensor where it belongs but the effort will be rewarded with better performance.

Among the useful tools for reducing false alarms are certain control panel programming options. Interior detectors in the path from the entry door to the alarm keypad should be programmed as "interior follower" zones. This will allow the homeowner to pass the detector during the exit delay after arming the system without triggering an alarm. It will likewise allow you to pass the detector after entering a delay zone (usually the door) upon return. However, if the detector is triggered before the entry door is opened the alarm is activated immediately.

Another false alarm reduction technigue that is available with some systems is called "zone ANDing". Two or more zones (sensing loops) are grouped together within an AND group. The alarm will only sound if at least two zones are tripped within a given time frame. Some panels, such as the Napco Gemini series, can be programmed to sound the siren on the first zone tripped but will only send a signal to the monitoring center after a second zone is tripped.

There is much more to false alarm reduction but the above are a few things you should consider when planning an installation.

I disagree. False alarm reduction techniques should be Course 101 for every user, installer, DIYer and service person who uses or works on an alarm system. The real tragedy of the security industry is the 98%+ ratio of false alarms which have caused police departments across the country to consider alarms and the alarm industry a public nuisance. Because so many alarms are a waste of time, police departments sometimes respond slowly or not at all to real alarm conditions. There is a far greater risk of tragey in this situation than arises from attempts to educate DIYers on the subject.

This is along the lines of what I was thinking about. I want to be SURE there's an incursion before the police are summoned. To me, at least, that implies that more than one type of sensor has been activated.

What concerns me the most is thunderstorms. We have terrible ones around here. Last night we had the power pop on and off at least ten times. Every system in the neighborhood that uses glass break detectors usually goes off. The alarm companies don't even bother dispatching anyone on really stormy nights. I'd like to figure out how to insure my system doesn't get triggered by lightning.

I agree with you, Bob, on this subject. The fewer false alarms, the better it is for every alarm system owner. As I said before, I had very few falses with the previous ADT system, and when the unit did trigger, the police were out here pretty damn fast. It can't hurt to educate DIY'ers on how to avoid false alarms.

-- Bobby G.

With the Gemini system that's a snap. You simply select which detectors will be ANDed and choose (or don't choose) to trigger the siren on the first AND zone trip.

I live in south Florida which allegedly gets more lightning per square mile per year than any other place on earth. I program glass break zones as interior followers which avoids the most common source of false alarms -- human activity. The IntelliSense (now part of Honeywell) detectors I use only occasionally false if the storm is intense and directly overhead. With a panel like the ELK M1Gold there's a way you could configure the glass breaks to be "lightning proof". When thunder triggers glass breaks it usually trips several of them. Have the glass break zones start a timer. If any glass break zone is tripped wait 10 seconds before triggering an alarm. If a second glass break activates within that time frame, take no action.

For the glass breaks on the HAI System, you could program a number of ways. In HAI, you can program a zone for away interior so they are only armed when no one is home. The HAI Firmware does not allow tinkering with the security functions, but you could write programs like you both describe. If one zone trips, and a second zone trips you could trigger an output to trigger the alarm. It is more of a complicated setup but you could accomplish your desired result of reducing false alarms in this method. You could also add timing like you have discussed.

Sounds clever. Wouldn't work if there's a bunch of people breaking in simultaneously. But if that's happening there are bigger problems like looting going on. I suppose interior motion sensing would cover it either way. Window rattling without interior motion is probably safe, window triggers with lots of motion and the voice prompt shouts "Lock and load!"

It is interesting to hear what different regional conditions require out of these systems.

Of course, human activity is what alarms are looking for in the first place, so whenever you start limiting the system, you're increasing the chances that a true alarm WON'T be reported.

The IntelliSense (now part of Honeywell) detectors I use

Usually, but not always in my experience. The trick here is to figure out what the weather conditions are independently of the intrusion sensors and then bias their reports accordingly. Now I'm not sure this means a using a lightning detector of some sort, or monitoring the NWS via the net (or better yet, by radio) or what, but clearly an "ultra smart" alarm system would take the weather into account.

We have a large apartment house nearby with a large, outside parking lot. It's bordered by open parkland and is the only structure above 10 stories for a radius of 10 miles. When there's a severe storm, a building hit will set off anywhere from 1 to 5 different car alarms. A smart thief could smash his way through a lot of cars on a night like that because the owners just shut their alarms off remotely, assuming it was the storm.

That's certainly one approach. It's based on a burglar only smashing one window, which is probably a pretty good assumption.

As a Floridian, you must be aware of what happened after Andrew. There were a *lot* of burglars smashing a *lot* of windows!

-- Bobby G.

I've seen it's capabilities. The HAI system is a far less configurable security controller than Napco's P9600 or ELK's M1Gold. It's a decent HA system and a usable security system but as an alarm it can't compare to the P9600 or the M1.

SIA has a CP-01 standard for false alarm reduction. The standard is likely to become a code requirement in many states due to the high incidence of false alarms from non-compliant systems. Has HAI implemented the standard? I was looking this up for someone else and I didn't see the Omni series on the list. I've asked ELK the same question and will post the answer when I receive it.

To my knowledge they have not integrated the standard, but with some programming knowledge you could do the same with HAI. Will the authorities accept it, that is anoter story...

In this case it's the activity of humans who live at the protected premises that causes most problems. I try to set systems up so that the homeowner is unlikely to accidentally trip an alarm. That usually means deactivating ("shunting") interior motion detectors and glass break detectors when the family is home until they retire for the night.

If only one triggers the alarm signal is sent. Only if multiple detectors trip in rapid sequence do you want to ignore glass breaks.

To effectively decide when to ignore a condition based on weather you would need a panel that simply does not yet exist. It's good to dream of ways to improve things in the future, but we have to deal with the present, using technology that is available today.

I don't really deal with car alarms. Most are useless junk.

Not one window -- one window within range of a single detector. Glass breaks are not designed to protect multiple rooms. Typically, there is one detector per room and each is (or really ought to be) on a separate zone (sensing circuit). I've yet to see a situation where the thief broke windows in separate rooms within a 10-second interval.

TTBOMK, noone was smashing lots of windows in different rooms of a single home in rapid succession. However, after Andrew hit there were large numbers of homes without power for weeks. None of their security systems were working. Also, phone lines were down so many alarms weren't able to call the central station.

You can't really build a commercially successful alarm or automation control that will handle earthquakes, hurricanes, forest fires or riots. It's also highly unlikely that we'll see an affordable HA system that makes processing decisions based on lightning proximity any time soon. For now, the best we can do is use the tools that are currently available, mix in a bit of ingenuity and experience and come up with something that minimizes problems while affording a reasonable degree of comfort and security.

They'll accept it if the panel is CP-01 compliant. The ELK M1 Gold is.

Don't know about you, but the glass breaks I use don't get tripped from thunder. Edmonton (Alberta) is also the "thunder & lightning" capital of the Prairies.

Zone "anding", timers... Tsk! This is a "poor tech's" method of covering up mistakes in placement, or the use of inferior equipment... It also very much enhances the perp's ability to defeat your system (accidentally or on purpose)...

And I suppose these glass breaks are in different rooms with windows facing in different directions... Good plan... But what if they're not?? What if the room is large enough to require two glass breaks?? What if the perp smashes a window that's close to or actually in an area where the coverage patterns happen to overlap?? So he smashes the glass and continues to break the broken bits out of the frame. There's your "first" trip, followed by a "second"... All within ten seconds... No problem... let's "ignore" whatever happens... Robert, your "experience" may be such that you would recognise when *not* to utilize this scheme. I think it's a huge disservice to even mention it though when most DIY's don't have the benefit of that "experience" and can't obtain it in the "two hours it takes you to teach someone how to professionally install on the phone"... Stick to providing a quality product, recommend they read the manufacturer's instructions and suggested installation tips, ensure they obtain the proper tester so that the unit can be set up correctly (in accordance with the manufacturer's instructions). Thieves will break in during lightning storms too, you know... ;-))

So, which non-falsing glass breaks DO you use? I'm sure I can give them a pretty thorough test. Unfortunately, as Marc H. and Bill K. have pointed out, some HA solutions are very location specific. I'll wager that another issue with glass breaks is how large a glass surface you are protecting, how well the glass is puttied into the frame, how much the windows resonate with ground and air vibrations, etc.

I'll admit, the timer trick was not a solution I would endorse whole-heartedly, although for a quick fix, it might cut down on falsing during storms. I really want something with the IQ to decide that it IS a storm, and that during storms it's going to take more than a single glass break to phone out to a central station. Surely the marriage of HA and security has matured to the point where that's not an impossible setup to construct. It may be as simple as monitoring the door intercoms at the front, back and side doors.

A thunderstrike will generate what I assume will be a loud "pop" on all intercoms simultaneously. That will tell me that there's either a storm or we're being carpet bombed. If that intercom noise spike coincided with a glass break signal, the system should disregard it. I'm just not sure of how to process those sound events into something I can feed into to HAI Omni LT to tell it "take it easy, it's only a storm." Reminds me of a black labrador we once owned that became so fritzed during bad storms that he would hide in the bathtub for hours. The vet finally prescribed doggy valium! Maybe there's a valium made for glass breaks. :-)

Good point. I don't really want to dumb down the alarm system, I want to make it smarter.

The problem with testers is that they really don't simulate the action they are trying to test for. But using them is certainly better than not testing at all.

-- Bobby G.

Well, that's certainly a common practice in the industry. Even so, I've triggered such systems inadvertently at least 4 or 5 times, once waking up an entire household. It's just too easy to forget about the alarm when letting the cat out or taking the garbage out or running after the FedEx guy. Even though those systems had an inside sounder to let you know you had just tripped the system, you can't hear it when outside trying to get that last bag of stinky kitchen garbage out to the curb as the trucks approach.

The point I was trying to make, however, is that burglary is a human activity, usually occurring from the outside to the inside. Perhaps the solution is to extend the alarm's sensing range to the lot perimeter. Unless the burglar parachutes in, he's going to have to cross ground before getting to the house.

I'll bet that will cover most of the falsing problems but we've had two hurricanes and a tornado pass through here in the last couple of years. Debris from such an even will trigger multiple glass break. Using the timer method would give a pretty erroneous reading of what's going on. The answer to me seems clear. Alarm systems really needs to know more about the current conditions than they do now. Falsing is a nationwide plague. Part of the reason is that commercial alarm panels are fairly stupid when it comes to determining whether an alarm is false or not.

I didn't expect to find such a panel off-the-shelf. But I'm pretty sure the sophisticated ones like the Elk and the HAI can be made more aware of their surroundings, even if it requires a separate HA controller and sensors like the Ocelot and their Bobcats to make the determination.

Take a look at Dr. Cheung's system. His house *knows* a lot more than any COTS solution will provide. It's doable.

IF rain sensors = wet AND all intercoms spike simultaneously AND the NWS download indicates "Stormy Weather" AND barometric pressure detector = "probable storm" THEN don't send alarm on glass break detection ALONE. I've seen TV weathermen show local maps of lightning hits. While I am not sure how I could obtain and reliable feed that info to my HA/alarm system, I'm betting something like that is possible. The question is, will it be reliable?

Just added that story as an aside. Getting an alarm company or the police to respond during a storm is probably going to be pretty hard just because they've experienced the parking lot/storm scenario a thousand times. When I helped the neighbor with the homicidal ex-husband she had an alarm installed right before a hurricane swept through. I camped out with her in the dark with my good friends, Mr. Smith and Mr. Wesson because had hammer-killer decided to come and kill her during the hurricane, no one would have noticed. The system did, in fact, go off and the alarm company never even called. That's when she called me and my little friends.

What - you've never seen "Scarface?" What if the army they sent to kill him came to the wrong house? :-)

After Andrew there was widespread looting and places where even the National Guard would not patrol because the residents out-armed them 20 to 1. I want to design an alarm system that's frugal enough with power to run for at least a week, even with the main AC power down. That can be partially achieved by using a high capacity, deep-cycle wheelchair battery. But it's going to require a lot of careful design and current measurements to be sure it will run that long.

Why do I want this? Because I've been without power for many days at a time around here. Even though I went to stay with relatives nearby, I really wasn't comfortable knowing that the 10 hour battery that came with the ADT system was already dead. I want something better than that kind of COTS crap to protect my home.

I wouldn't count on thieves to know that. I would count on them going down the street if a loud siren managed to sound once they forced my door open or smashed a window. That kind of protection is NOT outside the bounds of current technology. It's just going to take some engineering to design, assemble and test.

I've never been one to care much about what's commercially available if I know in my mind that it's technically feasible. While I don't see much payback in trying to design in earthquake alarms here in the NE, I DO see a very strong need to have an alarm system that can ride out the long-term power outages that occur here with increasing frequency and severity.

For the last 5 years, the local power company has slashed its tree-trimming budget. You can imagine the results. They've been forced by local authorities to put the trimmers back to work because a recent, rather mild storm knocked out thousands of area houses for up to a week.

I agree. But I also feel that if commercial panels are too stupid to know if it's thunderstorming outside, someone better learn them. Ask not for whom the bell tolls and the siren sounds . . .

-- Bobby G.

Ordinarily I ignore this poster due to his abusive behavior but the above needs to be answered for the benefit of other readers. First of all, Mr. Olson is not an installer nor even an experienced DIYer. Second, he resides in Vancouver, BC -- rather a long commute from Edmonton, Alberta. Third, there is NO acoustic glass break on the market that doesn't occasionally false from thunder. I've been installing glass break detectors for more than 20 years and I can assure the readers Mr. Olson is dead wrong.

Zone ANDing is one of the requirements of the current SIA standard for false alarm reduction. I suspect the gentleman on the standards committee would disagree with Mr. Olson as do I. However, the concept I explained in this thread could more accurately be called zone "NANDing" since it requires that one and only one acoustic glass break within a given group of detectors (say for example, all those facing the same direction) be triggered to generate an alarm.

Acoustic glass breaks can typically cover up to 50' (25' radius) so yes, they are usually installed in separate rooms. If you had actually installed any of them you would know that.

That is not essential but it can actually work to one's advantage using this technique. When windows in more than one room face the same direction, my method is enhanced because thunder will be more likely to trip several zones at once. Since most homes have more than one room on each side, that is usually the case.

If the room is longer than 50 feet on a side it might need two glass breaks. In that case, a simple modification to the rule will allow one to "OR" the two glass break zones within that room.

Quality acoustic glass breaks properly installed will not respond to breakage in an adjacent room. It is extremely rare to need more than one in a single room. However, if two or more were in a single room they can be zone ORed using a simple mod to the rule.

That's another misconception about glass breaks borne of inexperience. The good ones generally don't respond to secondary breakage once the window has been smashed. That has to do with the design of the glass break detector. Once the "skin" of the room has been broken, the additional noise generated by clearing the shards won't include sufficient low frequency sound to generate a secondary trip. Granted, if someone wanted to install the old Sentrol "ShatterBox" units or IEI junk from the 70's, that would be a problem. But those falsed when nothing at all was happening.

Robert,

He doesn't have any such glass breaks. He only posted so he could argue. A quick Google will be informative.

You are correct about the size of the pane. The larger the glass the more likely it will be to cause a false during thunder. How tightly it's puttied (or siliconed) isn't as much of a problem as one might suppose as long as the pane isn't shifting around in the frame. Room acoustices can be a significant factor in glass break performance. Small, tight spaces such as the entryway between inner and outer sets of doors in a commercial space can be very unfriendly to glass break detectors. Rooms with deep carpet, heavy drapes and upholstered furniture don't propagate sound as well as places with wood or tile floors, bare walls, etc.

Think it through a bit. If you understand glass breaks well and if you install them right this method can all but eliminate false alarms during thunderstorms without significantly reducing catch performance.

That would be nice but for the moment it's just not available at anything approaching a reasonable cost.

You can design and build a device to monitor microphones all around your own home easily enough. I have to deal with what is commercially available.

If you build it I can easily show you how to integrate it with the ELK-M1Gold. You're on your own with the HAI product. :^)

Not really. The other gentleman doesn't understand glass breaks at all. His "concerns" are based on fallacies about acoustic glass break sensors and a basic misunderstanding of what I proposed mixed with his oft-demonstrated personal animosity.

Smarter should include various means of reducing false alarms. You can make the system quite secure AND reasonably immune to false alarms. It requires thinking "outside the cabinet" at times. One thing I like about the ELK system is it does allow you tinker. You're not locked into a limited set of performance protocols based on a few zone types. You can make the system do what you need for each specific installation. (Note: I might be just a tad biased in this area since I'm an ELK dealer:))

That would be sad. Certainly there's variance in the quality of glass breaks. Maybe the best bet would be to get a roll of window foil and some varnish. Then again, maybe not. There's a reason we don't see thin foil bands around the edges of store windows anymore! Time = money!

But that's always what happens if the putty dries out. I rehab old houses and when they talk about rattling windows, I know the sound by heart. The last house I bought had pieces of plexiglas taped on the frame with packing tape. Oy! It was dirt cheap, however.

I suppose that the same sort of acoustic factors enter into whether a thunderclap gets amplified or attenuated.

So, which glass breaks do *you* recommend as being the most immune to falsing? If it makes any difference all the windows here are double-hung and casement style double-pane Andersons and the house is a small brick Cape Cod.

That's what I am trying to do - think it through. What possible tactics would a burglar use to break in? My hunch is that the rear basement windows are the most vulnerable simply because that's how my neighbors, left and right, were broken into. With crank-style basement windows there's an internal screen. Will that dampen the sound of the breaking glass?

Ah, but that's where I can afford to put sweat equity into something that a commerical dealer could not. I just have to ask myself if increasing the complexity if the system helps or hurts overall reliability.

Yes, in actuality, many of my CCTV cameras have built in audio that feeds back to a central panel in the basement. It wouldn't take much to take some samples during the storm and decided what level to use to make the determination "either thunder or WWIII is occurring." If I understood what I read (I think by Marc) the HAI panel will not be particularly useful in creating a "audio storm detector" because it does not have analog inputs. But I believe I can accomplish that part of the task with the CPU-XA. I simply need to send a signal to the HAI to indicate a possible storm in progress and a timer to reset that flag once no thunder has been heard for X-minutes.

Well, I assume that if Worthington recommends Elk sounders for HAI panels, that you can at least tell me what glass breaks to start with. Name a good enough price and I might even buy them from you (although Worthington is close enough to me for standard ground to arrive the next day).

oft-demonstrated

Well, I wish that everyone would tone down the personal bullshit. It seems that there's a grudge hierarchy in CHA going back at least 5 years. Someone should do an association diagram for newbies.

I'll admit my knowledge of glass breaks is very, very out-of-date. However, with a attempted murderer due to get out of jail in a my neighborhood eventually, decided I needed to install glass breaks no matter how much I don't like them.

Well, there's always the next house and I am not really adverse to having two entirely independent alarm systems installed, especially if I can put in the two for less than the price of a crappy commercial one.

That might be the best way to eliminate falsing. Two independent systems have to agree there's an intrusion before making any phone calls.

-- Bobby G.

As long as you don't need to do numerical computations on the analog inputs, and simply use the controller's software to detect a threshold (i.e., a specific value that when exceeded triggers an effect; "IF volt > x THEN etc ") you can accomplish the same thing in hardware _externally_ with a simple voltage divider formed by two fixed resistors or a pot and an opamp (or transistor, comparator or TTL or CMOS digital IC) to shift the level to the desired on/Off values (0 or 5vdc; 0 or 12vdc etc.). The usefulness of the analog input depend in large part on the software/firmware of the controller.

Well there are surely many ways. For example CyberHouse has a client called "Listen" that can be trained to listen to an audio signal and alert the CyberHouse server when that pre-identified audio signal is heard. This can be a washing machine, noisy water valve, doorbell, Trouble is that thunder is not a single sound

shows some sonograms of thunder that confirm our qualitative assessment that thunder has different so it may be difficult to train the software to 'hear' lightning. (I dunno.)

But because because thunder is caused by lightning, it might be practical to detect the lighting that causes the thunder and simply disregard glass breaks when lightning is indicated in your area on the www or if your lightning detector goes off.

The topic of lightning detection has come up before in the context of powerline surges/outsges in this newsgroup (Goggle).

Marc Marc_F_Hult